During World War II, the US Army was interested in developing a system that rapidly detects bacterial biowarfare agents in aerosols. Gucker and O’Konski (1947) set up and developed a photoelectronic counter which injected an air stream containing particles into a sheath stream on a dark-field microscope focal point. This first air-fluxes cytometer detected 0.5 micron events by scattered light. In 1942, Coons et al. used, for the first time, an anthracene-associated antibody (UV-excitable fluorochrome) to detect Streptococcus pneumoniae. One decade later, Coons and Kaplan (1950) described the first fluorescein isothiocyanate (FITC) antibody association. In 1945, Coulter developed a process that permits counting events and measuring cell size by conductance variation. In this primary system where cells pass one by one into a very thin capillary and cross a light beam, extinction number corresponded to the cell number. However this technique led to many clogs. In 1949, Coulter precisely measured cell volume in a saline suspension based on electrical impedance variation proportionality due to the fact that sheath fluid and cell suspension are ionic liquids and cells surrounded by a lipid membrane are poor conductors compared to saline fluid. Coulter’s cell counters evolved into cell analyzers and were rapidly adopted by clinical laboratories for blood cell counting (Brecher et al. 1956; Mattern et al. 1957).

In 1953, Crosland-Taylor used the laminar coaxial flux properties described by Reynolds in 1883, to develop a device counting red blood cells suspended in a sheath fluid through a capillary (Crosland-Taylor 1953). Simultaneously, photomultiplier tubes appeared, increasing optic signal precision. In the middle of the 1960s, Fulwyler described the electronic separation of cells by volume (Fulwyler 1965). Concurrently, Katmentsky et al. (1965) described a rapid cell spectrophotometer, which made rapid absorption measures. In 1965, he also studied optical properties of cells passing across a laser beam, such as scattered light intensity (forward scatter) and UV-light absorption (DNA quantity) after staining. In 1967, he developed a spectrophotometric cell sorter (Kamentsky and Melamed 1967). In the late 60s, Göhde’s Partec (Münster, Germany) developed an analyzer (Dittrich and Gohde 1969) built around a Zeiss fluorescent microscope for DNA analysis with ethidium bromide (LePecq and Paoletti 1967). One year later, Phywe (Göttingen, Germany) commercialized it for the first time under the name impulsecytophotometer ICP11.

In 1970, Kamentsky founded Bio/Physics Systems (Mahopac, NY, USA) and in 1971 sold the Cytograph and Cytofluorograph systems. For the first time, Van Dilla et al. (1969) equipped these analyzers with respectively 633 nm He–Ne laser and 488 nm argon ion lasers. Excitation by a laser instead of a lamp allows optimal beam focusing on cells. Cytofluorograph allowed the measurement of forward scatter light, green and red fluorescence emission (530 and 640 nm respectively). In 1972, Bonner et al. developed a new system named fluorescence activatedcell sorting (FACS) which purified sub-populations (Bonner et al. 1972) and was equipped with a water-cooled argon laser to analyze FITC- and Rhodamin-coupled antibody fluorescence. To develop cell sorting, Fulwyler (1965) and Kamentsky and Melamed (1967) used a droplet deflection system developed by Sweet (1965) at Stanford University (Stanford, CA, USA) for ink jet printers in 1965. With this droplet deflection technology, the saline sample stream was broken into droplets containing cells. Those of interest, with selected measurement values, were electrically charged at the droplet break-off point and droplets were then deflected by the means of an electric field into a collection tube. Their larger cell sorter rapidly measured scattered lights at several angles (Mullaney et al. 1969; Salzman et al. 1975a, 1975b; Steinkamp et al. 1973), fluorescence and measured cell volume by electrical impedance variation. In 1973, Hulett et al. developed a rapid cell sorter (Hulett et al. 1973). In 1974, the first commercial flow cytometric differential counter was the Hemalog-D (Technicon, Tarrytown, NY, USA), which led to the establishment of a white blood cell classification (Mansberg et al. 1974; Ornstein and Ansley 1974). The same year, Becton–Dickinson (now BD Biosciences, San Jose, CA, USA) commercialized a second cell sorter version named FACS II, which measured FSC and 530 nm emission fluorescence. Simultaneously, the Max Planck Institute (Göttingen, Germany) built a multiparameter cytometer cell sorter (Arndt-Jovin and Jovin 1974), and the Partec/Phywe society (Göttingen, Germany) commercialized the Impulsecytophotometer ICP 22. In 1975, Partec commercialized the two-parameter Particle Analyzing System PAS™ 8000. In the middle of the decade, Coulter Electronics (now Beckman Coulter, Fullerton, CA, USA), known for hematologic counters, made their first cell sorter named two-parameter-sorter-1 (TPS-1) with an air-cooled 35 mW argon ion laser, measuring FSC and one fluorescence. Several excitation wavelengths were introduced to flow cytometry: the first analyzer (Curbelo et al. 1976) used five illuminating beams from a single arc lamp; the second named Cytomat-R (Shapiro et al. 1977) with three laser beams was built by Shapiro in order to analyze FSC, SSC and several fluorescence colors (Shapiro et al. 1976) allowing morphological gating (Shapiro 1977). Both could detect up to 30,000 events/s. In 1977/78, Coulter Electronics developed the EPICS (Electronically Programmable Individual Cell Sorter) system with a 5 watts argon laser and data multiparametric analysis. In 1978, at the Conference of the American Engineering Foundation in Pensacola, Florida, the cytophotometry technique was renamed flow cytometry, a term that quickly became popular. The same year, the Society for Analytical Cytology (later renamed International Society for Analytical Cytology and then International Society for Advancement of Cytometry) that edits the journal Cytometry was created. In 1978, Schlossman started the production of monoclonal antibodies directed against blood lymphoid antigens. New fluorochromes for multicolor flow cytometry were also developed (Reinherz et al. 1979). In 1979, Partec commercialized the PAS-II instrument. At the same time, NIH scientists added a 568 nm krypton laser to the flow cytometer and after a developmental phase, FACS IV was the first dual-laser cytometer created, commercialized by Becton–Dickinson (Steinkamp et al. 1979). Coulter and Ortho (that bought Bio/Physics Systems) manufactured flow cytometers that measured FSC, SSC and fluorescence, and analyzed several thousand events/s. After successfully performing a dual-color immunofluorescence experiment, Loken et al. (1977) introduced fluorescence compensation in the process.

In the early 1980s, optical emission systems appeared. In the mid-1980s, Lawrence Livermore National Laboratory (LLNL) created the first high speed cell sorter prototype (Peters et al. 1985) for human chromosome separation (Gray et al. 1987). This high-speed cell sorter could sort 20,000 cells/s at 200 psi (pound per square inch) pressure, three times faster than conventional sorters (8,000 cells/s at 12 psi). In 1985, cell analyzers allowed up to three color analysis, such as EPICS C analyzer (Leif et al. 1985) (Coulter) with an arc lamp source or FACScan™ (Becton–Dickinson) with a 15 mW air-cooled argon laser source. Competition started between the different flow cytometer manufacturers concerning the simultaneous analysis capacity of fluorescence. In 1987, Partec introduced Cell Analyzer CA-II. In 1989, LLNL produced a second generation of high speed cell sorters allowing 200,000 events/s sorting at less than 100 psi pressure, with parallel organization of the data analysis process (van den Engh and Stokdijk 1989).

In 1990, flow cytometer capacity increased to measure seven fluorescences simultaneously. In 1991, Partec commercialized the two Particle Analyzing Systems PAS-III and PAS-IV. In 1994, Cytomation (Dako/Cytomation and now Beckman Coulter) commercialized the MoFlo^®MLS flow cytometer developed by van den Engh, the first high-performance cell sorter for high-speed sorting applications. From the mid-90s, smaller diode and solid-state lasers were more and more often incorporated into flow cytometers. Because many objects are too large and too fragile for conventional flow cytometry, in 1998, the first Complex Object Parametric Analyzer Sorter (COPAS™) was introduced. Over the next decade, Union Biometrica expanded the COPAS platform into a family of fully automated systems for high throughput analysis, sorting and dispensing of large objects ranging from 20 to 1,500 microns.

In 2000, Apogee Flow Systems Ltd sold systems dedicated to environmental bacteria detection. The US Army chose the A40military flow cytometer (Apogee Flow Systems) for its high sensitivity to small particles. In 2001, a violet diode laser was used for the first time on a flow cytometer (Shapiro and Perlmutter 2001). The same year, the BD LSR™ II flow cytometer proposed detection of fourteen fluorescences, a powerful tool for elucidation of the complex immune system (De Rosa and Roederer 2001; De Rosa et al. 2001). One year later, in 2002, the FACSAria™ (BD Biosciences) became the first cell sorter with a fixed optical emission system. Partec inserted a flow cytometry laboratory into a french car: CyLab™ was the first mobile flow cytometry lab. Partec later proposed the CyLab™ Plus, an advanced mobile HIV monitoring laboratory based on a 4-wheeldrive transporter vehicle using CyFlow^® technology which can also run on a car battery (12 V DC power). In 2004, a Partec’s CyFlow^® integrated the International Space Station (ISS). The same year, new nanocrystal semi-conductor fluorochromes named Qdots^® appeared. All five different nanocrystals are excited with the same long-wavelength UV lamp and their size determines their color. With Qdot^® technology, Perfetto et al. (2004) realized seventeen different fluorescences per cell experiment. In 2007, van den Engh developed a new high-speed cell sorter named Influx™ (Cytopeia) and sold his society to BD Biosciences while Beckman Coulter acquired Dako/Cytomation.

In 2010, Sony, Merck Millipore and Danaher respectively acquired iCyt, Guava Technologies and Beckman Coulter. New ergonomic and compact analyzers appeared: Cube 8 (Partec), easyCyte™ 8HT (Merck Millipore) and the first acoustic focusing (Ward et al. 2009) cytometer Attune^® (Applied Biosystems™ by Life Technologies™). Furthermore, new analyzers appeared LSRFortessa™ (BD Biosciences), Gallios™/Navios™ (Beckman Coulter) and a high-speed cell sorter Astrios™ MoFlo^® (Beckman Coulter). This year, in 2011, BD Biosciences acquired Accuri Cytometers and commercialized a new benchtop analyzer, the FACSVerse™. Two new bench-top analyzers appeared MACSQuant^®VYB (Miltenyi Biotec) and Auto40 (Apogee Flow systems). Partec produces the CyFlow^® Cube Sorter, a new bench-top mechanical cell sorter.

For flow cytometry acquisition, cells need to be in suspension in a tube or a plate. Some flow cytometers are equipped with a carousel or a rack loader (16, 30, 32 or 40 tubes capacity) containing a bar-code identification option and a vortex mixing system prior to sample aspiration. Some analyzers are equipped with a high throughput 96 or 384-well plate loader or a built-in sample tray for a 96-well microplate and/or tubes. Some systems are temperature-controlled, whereas others offer the possibility to add reagents or drugs during analysis.

Cells pass one by one across laser(s) beam(s) for individual analysis. This hydrodynamic focusing (Fig. 1 part 1) is based on coaxial laminar flow dynamic properties described in 1883 by Reynolds. To ensure a good quality of hydrodynamic focusing, the fluidic system must be very stable. The sample suspension is first pressurized and then injected into a sheath core flow before passing thought a nozzle. The cellular suspension speed is dependent on the sheath fluid pressure that is fixed for an analyzer and adjustable for a cell-sorter. The higher the sample pressure the more cells have an opportunity to move laterally in the stream, causing a decrease in precision of hydrodynamic focusing, and consequently a drop in quality of the sample analysis, due to increases in the Coefficient of Variation of the peak values. The nozzle design is essential for obtaining laminar fluxes (Fulwyler 1977) in order to transport cells to the center of the stream: nozzle aperture should be narrow and perfectly calibrated (between 50 and 2,000 microns, depending on applications). This hydrodynamic focusing results from several stability developments (Steen 1990), and is used in most flow cytometers. However, other focus concepts exist, such as microcapillary technology or acoustic focusing (Ward et al. 2009) which allows alignment of cells by sound waves (Goddard et al. 2006, 2007; Goddard and Kaduchak 2005) without damage to cells (as ultrasound for fetus in utero medical exam).

Flow cell could be qualified as the heart of the system because it is the place where laser beam and cells interact (Fig. 1 part 2). That is the starting point of scattered lights (forward scatter and side scatter) and fluorescence signal emissions detected with an optical collection system before amplification and electronic digitalization. This intersection between excitation light source and cells is occurring in a quartz chamber or in the air at the nozzle exit. Due to the important variation of refraction indices between air and sheath fluid, the reflection phenomenon implies lack of light signal so the amount of emission wavelength collected in the air is less important than in quartz chambers; numerical aperture of the objective and sheath fluid pressure are important in this case. In contrast to quartz chambers, optical systems in the air have to be aligned daily (Watson 1999) to ensure sensitivity and quality.

Cell illumination with an excitation light source provides scattered (forward scatter and side scatter) and fluorescent lights. The first flow cytometers were equipped with a mercury arc lamp, but light focalization on a small analysis area and light emissions collection were very poor, unlike those delivered by a coherent and monochromatic laser beam, direct or fibered through prisms and optical lens (Van Dilla et al. 1969) also used to control illumination point geometry. Indeed, contrary to circle geometry, the elliptic and now the rectangular geometry laser beam homogeneously illuminate cells whatever the lateral cell position in sheath fluid. The variable power intensity laser (15–200 mW) is function of the flow cell (quartz or jet-in-air). Emission fluorescence is quantitatively proportional to excitation intensity of the fluorochrome. At the present time, excitation light sources are systematically diode lasers or lasers, except for one analyzer which uses mercury arc-lamp for UV-excitation. Since 2001, violet laser diodes are used as light sources for cytometry (Shapiro and Perlmutter 2001). Use of several excitation sources increases the number of fluorochromes detectable simultaneously and directly impacts on the number of cellular parameters correlated (Crissman and Steinkamp 1982; Perfetto et al. 2004). Some flow cytometers were equipped with four different excitation sources (Greimers et al. 1996) or more, today cell sorters can be equipped with up to ten lasers. Laser co-linearity (several excitation sources in the same plan) implies alignment simplicity and stability but involves more fluorescence splitting than in a laser non-co-linearity system (one plan by excitation sources), which allows effective fluorescence splitting, but creates a spatial gap, electronically converted to a temporal gap (time laser delay) in order to synchronize several light signals from one cell.

When cell pass through the excitation source, the laser beam is refracted in all directions. Light diffusion at small angles (forward scatter light, FSC, FALS) is collected in the axis of the laser beam by a photodiode or a photomultiplier tube; the magnitude of the forward scatter light is correlated to the relative-size of the cells. A shutter bar masks the signal due to the laser beam on the detector. Some analyzers permit a choice between 1–8° and 1–19° collection of FSC signal (field stop) to detect, respectively, only large events or total events. Light diffusion at large angles (side scatter light, SSC, WALS) is collected at 90° of the laser beam like fluorescence emission lights (Fig. 1 part 3). Orthogonal side scatter light is a combination of diffusion, reflection and refraction caused by structural complexity into the cell. In flow cytometry, diffusion phenomena are very complex (Salzman et al. 1979). Side scatter and fluorescent lights are filtered by dichroic mirrors and adequate emission filters (band pass, short pass or long pass) (Steinkamp et al. 1974) into an emission optical system which directs lights at different wavelengths towards appropriate detectors. The role of the dichroic mirror is fundamental because it selects, reflects and transmits light signal band pass towards specific detectors collecting specific fluorescence. Light reflected by dichroic mirror is better than transmitted light, which originates in a multicolor staining, which signal passes through several filters and loses a significant part of its initial intensity. Thus, architecture of the emission optical system is essential to minimize losses and usually the user has the possibility of changing optical filters for each fluorescent measurement.

Detectors (photodiode and photomultiplier tubes) convert and amplify light signals (photons) from cell passage into a laser beam to an electric signal. This pulse is then digitalized, recorded and treated by specific software. Photodiodes are used for high intensity signals and photomultiplier tubes (PMT) for low intensity signals that need to be amplified with dynodes succession into PMT. Some analyzers also allow increasing gain, generating significant unspecific signals (background noise) so that is preferable to first increase PMT voltage before gain for low intensity signals.

An impulsion’s characteristics are pulse height or peak value, pulse width and pulse area or integral value. Impulsion length is also named flight time or collection integration time. Trigger is a parameter chosen by the user based on a discrimination value (threshold), often FSC, below which events are not considered by the electronic system.

The flow cytometer electronics (Fig. 1 part 4) measures several thousand events per second, limited however by electronic dead time (Snow 2004), which is today less than 5 μs. Most efficient cell sorters have no dead time, and thus no hard aborts. Analyzers digitalize signal with analogic to digital converter (ADC) by converting voltage value to digital value and determining the channel number: 256 (8 bits), 1,024 (10 bits), etc. on logarithmic or linear scale (Fig. 1 part 5). Some analogical and all new digital flow cytometers allow retreatment of the data and modifications of fluorescence compensation matrix post-acquisition. Flow cytometers simultaneously analyze pulse height, width and area for each cell allowing doublet-exclusion. Actually, the highest resolution is 32 bits (2³² channels), divided on a five decades logarithm scale (up to seven decades for some). Most recent cytometers offer a biexponential scale under the axis. Events visualization is characterized by exhibiting pseudo-linear like behavior for values near zero, and transitioning it to a pseudo-logarithmic behavior values distant from zero (Novo and Wood 2008; Parks et al. 2006). Thus, events with lower fluorescence intensities are grouped and cell populations appear homogeneous.

Cell sorters offer the possibility of isolating subpopulations of cells of interest with high recovery and high degree of purity from heterogeneous cell mixtures based on light scattering and fluorescent characteristics.

Two kinds of sorting mechanisms can be described (Ashcroft and Lopez 2000).

First, the mechanical sorter employs a mechanical catcher tube to sort cells of interest. The catcher tube is located in the upper portion of the flow cell and moves into the stream to collect the cells. If the cell is identified as a cell of interest, it is captured by the catcher tube and collected into a tube or a concentration module; otherwise it is dispatched to the waste tank. Ideal for pathogenic samples, this method produces no biohazardous aerosols, but is considerably slower in sorting rate (300 cells/s with BD FACSCalibur™ and CyFlow^® Sorter module by Partec). The mechanical system permits sorting of only one population. These flow cytometers are designed with a closed fluid system, thereby making them ideally suited for working with potentially biohazardous samples.

Second, most high-speed cell sorters use electrostatic deflection of droplets (“jet-in-air” method) (Chapman 2000). The advantage of this system is the possibility to sort from one to six sub-populations of cells (depending on cell sorters) simultaneously at low- or high-speed. As sorting generates aerosols in the chamber (Schmid and Dean 1997; Schmid et al. 1997) working with potentially biohazardous samples (Perfetto et al. 2003) will require implementation of appropriate precautions.

The cell suspension is directed into a stream, which emerges from a vibrating nozzle and breaks up into individual droplets (Fig. 1 part 6). The nozzle vibration conditioned by the drop drive frequency (ddf) is the number of drops generated per second proportionally to amplitude level. The fluidic system stability allows a good evaluation of the cell localization inside the droplets (Petersen and van den Engh 2003) if all conditions are accomplished. A droplet containing a cell of interest is positively or negatively charged and goes through an electric field between two deflection plates before being deflected into collection tubes (Fig. 1 part 7).

Several parameters should be considered for cell sorting optimization. According to type of cells, it is necessary to adjust sheath fluid pressure linked to the orifice diameter of the nozzle tip; the number of drops formed per second depends upon nozzle vibration (frequency and amplitude). The design specification for a cell sorting instrument also includes set up of break-off point (point where the stream breaks into droplets), drop delay (distance between the laser beam interception of the cell and the break-off point), and purity level (sort mode’s choice). These calculations can be automatic.

Thus it is possible to sort from 1 to 6 ways simultaneously into several containers (0.6, 1.5, 2.0, 5.0, 15 or 50 mL tubes, multi-well plates until 1,536, slides).